Regarding the sterileness observed when comparing some of these drivers, I think the issue here can be understood most simply by examining and comparing the power rating specs of driver as related to magnet strength (and also observing damping characteristics) -- the AXI 2050 has a 150w nominal 300w continuous power rating with no magnet power specified (BL Factor, Maxwell, Flux Density); Radian 951BePB TAD killer...: 125w aes 250w continuous (how extreme) 1.9t flux density; Beyma CP755ND: 60w aes 120w continuous 2.2t flux density with a larger magnet than any of its contemporary counterparts to my knowledge; TAD TD4001 30w aes, 60w continuous, 2t flux density, 228000 maxwell magnet strength; Altec GPA 288 (20w aes?) 40w continuous, 2t flux density; Goto Unit SG2880BL 5w rating with a beryllium diaphragm, 2.4t flux density 2,250,000 maxwell magnet... The point being while a new driver may be super clean in what it's able to render and will also play loud without distorting or being damaged, a mega magnet drivers the Goto being the most substantial example, with super sensitive low mass diaphragms should yield far greater detail, nuance and musicality. The Radian for example and Trueextent diaphragms are by no means a replacement of the TAD TD4001s and the likes. One is for mass market commercial applications, the other is a cost no object lab grade instrument advanced in the way that an ion thruster is. A lighter weight diaphragm coupled with a more powerful magnet would likely yield greater frequency extension in this instance.
Last edited:
Did you use the same passive crossover for all compression drivers?
I have to repeat the same question as TNT: no EQ applied to the different combinations?
No EQ at this point. The crossover I'm using is a Beyma FD212 with different attenuation settings and a 3.5khz 3db attenuation point. Some of the drivers were inherently shrill. I rejected those with the belief that added eq would be an unnecessary complication of an inferior product. Additionally certain drivers felt more alive, nuanced and interesting to listen to regardless of volume. The CP755ND, Altec 288, and BMS 4552, while all quite different had a unique characteristic that I found impressive and justified keeping. I'm a minimalist and the goal with this design is to achieve a most satisfactory sound while incorporating the fewest variables possible. While I've never heard either, a Danley synergy horn to me is overkill and I question the choice of compression driver with regards treble performance, an AER Pnoe is too minimal I think, in that I presume the back loaded horn is to achieve a semblance of bass performance at the cost of coherence, and it likely beams and lacks in treble performance. I think a properly implemented two way is a good balance. Though I'm sure the urge to explore more complex designs will persist.
I contemplated purchasing a pair of Meyersound Acheron Designers which feature a JBL compression driver (the 2450j I believe) and have heard the Blue Horns (same driver). While technically good, they simply didn't have the magical quality that I seek. With a good set of speakers, the sound should feel infinitely deep. Many factors can reduce this quality including lack of extension, overly damp drivers, over reliance on DSP, overly damp rooms, limited amplification gain, limited amplification power, limited dynamic range, etc. to a point where the sound goes from having infinite depth and nuance to sounding squashed and finite. Some of these factors effected my opinion of the Blue Horns. I own a pair of Genelec 8331s and have compared them with a pair of 8351s in my living room. They're great given how compact they are though they are not as satisfying to listen to as my living room setup. They sound squashed and minuscule in comparison, despite the rudimentary nature of my explorations thus far.
I have three amplifiers -- a Topping LA90, a 211 tube amp and a FU9 amplifier. In addition I have a MiniDSP Flex which I intend to start using. Being somewhat lazy I also ordered a K231 active crossover with some different crossover points to experiment with and compare.
I contemplated purchasing a pair of Meyersound Acheron Designers which feature a JBL compression driver (the 2450j I believe) and have heard the Blue Horns (same driver). While technically good, they simply didn't have the magical quality that I seek. With a good set of speakers, the sound should feel infinitely deep. Many factors can reduce this quality including lack of extension, overly damp drivers, over reliance on DSP, overly damp rooms, limited amplification gain, limited amplification power, limited dynamic range, etc. to a point where the sound goes from having infinite depth and nuance to sounding squashed and finite. Some of these factors effected my opinion of the Blue Horns. I own a pair of Genelec 8331s and have compared them with a pair of 8351s in my living room. They're great given how compact they are though they are not as satisfying to listen to as my living room setup. They sound squashed and minuscule in comparison, despite the rudimentary nature of my explorations thus far.
I have three amplifiers -- a Topping LA90, a 211 tube amp and a FU9 amplifier. In addition I have a MiniDSP Flex which I intend to start using. Being somewhat lazy I also ordered a K231 active crossover with some different crossover points to experiment with and compare.
Last edited:
While this is valid approach for your application that uses no EQ everyone should know that the frequency response by far dominates all other effects that affect perceived sound quality, like distortion with home listening levels, so perhaps the badly deemed drivers are just fine with systems that has been EQ:d.
While it is an indication for good engineering the response be nice from get go but there might be better compromises in the engineering made with assumption that the response is EQ:d in the application, like its usually done on live sound where most of the products are geared towards I assume. There is nothing to be afraid of with EQ, quite contrary it can make the sound better. If system has no EQ then it makes sense to pick the best sounding one as again, the other aspects an engineer might have chosen to compromise the response are less important in home application with low watt use.
While it is an indication for good engineering the response be nice from get go but there might be better compromises in the engineering made with assumption that the response is EQ:d in the application, like its usually done on live sound where most of the products are geared towards I assume. There is nothing to be afraid of with EQ, quite contrary it can make the sound better. If system has no EQ then it makes sense to pick the best sounding one as again, the other aspects an engineer might have chosen to compromise the response are less important in home application with low watt use.
Last edited:
That was my finding too with regard to frequency response. In the interest of building a two way system I dismissed the drivers which don't lend themselves to that design. I've spoken with three people who have built systems using AXI2050s and in every instances it's agreed that the treble response is lacking. One member was using a K402 horn with a Celestion compression tweeter swapped for a Beyma TPL200, another said that with as much as 10-15db equalization in the treble region it still wasn't there. The Altec 288s play similarly low yet also have great extension. Supposedly the Yamaha JA6681B is an upgrade to the 288 and the diaphragm has an interesting low mass surround however the driver features a ferrite magnet disguised to resemble an Alnico (kind of like the Celestions in my opinion) which seems ridiculous.
The HF1440s sound shrill. The HF206s aren't extended enough and sound dull. The HF108s and 4552s are very nice. The rest are mediocre and or have fit and finish issues.
The HF1440s sound shrill. The HF206s aren't extended enough and sound dull. The HF108s and 4552s are very nice. The rest are mediocre and or have fit and finish issues.
Last edited:
Crossover should be designed to maximize good sides of the driver and to minimize the bad ones, to make good integration of driver to the woofer, to EQ (to some extent) frequency irregularities of the driver, and to enable smooth of-axis response around the crossover point. Clearly, it is not possible one and same crossover to be optimal for two different drivers, not to mention a wide variety of drivers. Therefore, your approach (using the same crossover for different drivers) is wrong and have zero value, but unfortunately, that crime is not covered by federal laws and you will escape suing. 
With optimal crossover for each driver your driver preference will be totally different.
With optimal crossover for each driver your driver preference will be totally different.
How can you tell anything worthwhile in this application? The published charts for all the drivers are done on different flares, sometimes using different methodologies so using them as some sort of referent when the drivers are used sans flare means nothing. Likewise each should be xovered and EQd individually to get them flat or to whatever curve, but the same for each driver, so they are on the same playing field.
The drivers which I promptly rejected sounded smeared, shrill, undefined, dull etc. relative to my reference standard which I've specified previously. A horn does not resolve said issues nor should one have to rely on eq to do so. The one thing that I do think could help in certain instances which was observed by Gedlee to deal with issues associated with beaming and shrillness to an extent is open cell foam. That stuff is interesting in this type of application and I intend to do some experimenting with it down the road. The HF1440s for example simply don't sound very good in my experience. I don't like the idea of having to correct something that is poorly implemented in my opinion, or not as well implemented as another available option. The B&C drivers which I tested are a lesser driver in every way on paper to the Beymas, my listening experience reaffirms this.
Regarding different horns being used for testing and to derive measurements I am aware and have found that they're all within a fairly standard window or spectrum of geometry types -- nothing exotic or out of the ordinary. In some instances a plane wave tube measurement is also presented. I have/had the majority of basic horn varieties on hand and mixed and matched in pretty much every instance of testing matched and unmatched drivers. I found that more horn loading equated to more coloration. The argument of increased sensitivity due to reduced air mass at the diaphragm is not a justifiable argument for using a horn in my experiences thus far. The A290 horn is the largest one I have and it yields good results. Horns that feature a long narrow throat seem to smear the content and reduce high frequency definition, compared with a wider dispersion device. This is certainly an issue that is resolved to a degree with proper matching of entry and exit angles. The Altec 288s for example are clearly made for a highly directive long throw horn where as the CP755ND likely less so. The JBL M2 horn seems like a nice middle ground for a contemporary driver with a shallow throat and wider exit angle. Those Goto alphorn looking horns for example seem ridiculous -- the intent being absolute maximum sensitivity which I'm sure they may achieve though also possibly result in a reduction neutrality at best.
Regarding different horns being used for testing and to derive measurements I am aware and have found that they're all within a fairly standard window or spectrum of geometry types -- nothing exotic or out of the ordinary. In some instances a plane wave tube measurement is also presented. I have/had the majority of basic horn varieties on hand and mixed and matched in pretty much every instance of testing matched and unmatched drivers. I found that more horn loading equated to more coloration. The argument of increased sensitivity due to reduced air mass at the diaphragm is not a justifiable argument for using a horn in my experiences thus far. The A290 horn is the largest one I have and it yields good results. Horns that feature a long narrow throat seem to smear the content and reduce high frequency definition, compared with a wider dispersion device. This is certainly an issue that is resolved to a degree with proper matching of entry and exit angles. The Altec 288s for example are clearly made for a highly directive long throw horn where as the CP755ND likely less so. The JBL M2 horn seems like a nice middle ground for a contemporary driver with a shallow throat and wider exit angle. Those Goto alphorn looking horns for example seem ridiculous -- the intent being absolute maximum sensitivity which I'm sure they may achieve though also possibly result in a reduction neutrality at best.
I am running a pair of Yuichi A-290 with Radian NEO 950 PB 16 drivers, not the Be version which are pretty expensive. Compared to some of you guys I don't have a lot of experience, still. I have found EQ mandatory with these horns and all predecessors regardless of driver used with them. With proper EQ the A290 Radian combo is very good.
Some manufacturers are now using 90 x 40 horns measured at 1 meter out when they are specifying their drivers, others use the plane wave tube, some both. What is obvious from looking at the response curves is that they all need EQ to varying degrees (regardless of how they were measured).
Brett's point is an important one, without proper EQ to the same target curve using the same horn how can you make a valid comparison? On good drivers I would expect the differences in frequency response to swamp differences due to distortion, etc.
Last time I checked TAD used TruExtent diaphragms in their drivers. My understanding is that the magic is in the phase plug. I have several friends who use TD-4001 or TD-2001, they are very good - I just can't justify the money and have not tried them here..
Some manufacturers are now using 90 x 40 horns measured at 1 meter out when they are specifying their drivers, others use the plane wave tube, some both. What is obvious from looking at the response curves is that they all need EQ to varying degrees (regardless of how they were measured).
Brett's point is an important one, without proper EQ to the same target curve using the same horn how can you make a valid comparison? On good drivers I would expect the differences in frequency response to swamp differences due to distortion, etc.
Last time I checked TAD used TruExtent diaphragms in their drivers. My understanding is that the magic is in the phase plug. I have several friends who use TD-4001 or TD-2001, they are very good - I just can't justify the money and have not tried them here..
Last edited:
Yeah, you do. You're also testing them in an unusual manner as the CDs will expect to be loaded on the horn side so the diaphragms may not be centred in the magnetic field as designed.
It's easy to just dismiss, but some of us have been doing this a long time. I've never found a CD used without a horn/WG to sound in any way representative of one in the correct horn/WG with EQ applied, directivity matched etc..
What is your eq like for your Radians? The Truextent TADs are for the American audience. It's possible that some vendors replace the diaphragms and sell TAD diaphragms independently. The vapor deposition beryllium TAD diaphragms are 5000 a pair.
The CP755ND is a more robust driver than most others which I tested. The result is better frequency extension. The response curve on the datasheet is obviously smoothed. I imagine it's pretty jagged near 20khz as titanium starts to break up substantially around 10khz. This is not reflected in the sound that they produce however. I plan to measure things and do some equalization to resolve distortion related abnormalities. I prefer a driver thats inherently extended and relatively flat to one which requires a HF boost. I'm sure I'll revisit some new drivers once I start making measurements and experimenting with eq -- likely drivers with mylar diaphragms and low power handling.
The CP755NDs simply yielded more detail compared with anything else I tested including tweeters and three way configurations. I attribute this in part to them being more extended however when comparing with the more sensitive BMS 4552 for example as a two way and three way with the Celestion AXI 2050 the Beymas were again more resolving to a great degree. Same observation applies to every single other configuration tested.
The CP755ND is a more robust driver than most others which I tested. The result is better frequency extension. The response curve on the datasheet is obviously smoothed. I imagine it's pretty jagged near 20khz as titanium starts to break up substantially around 10khz. This is not reflected in the sound that they produce however. I plan to measure things and do some equalization to resolve distortion related abnormalities. I prefer a driver thats inherently extended and relatively flat to one which requires a HF boost. I'm sure I'll revisit some new drivers once I start making measurements and experimenting with eq -- likely drivers with mylar diaphragms and low power handling.
The CP755NDs simply yielded more detail compared with anything else I tested including tweeters and three way configurations. I attribute this in part to them being more extended however when comparing with the more sensitive BMS 4552 for example as a two way and three way with the Celestion AXI 2050 the Beymas were again more resolving to a great degree. Same observation applies to every single other configuration tested.
Last edited:
I didn't know they used the TreExtents. I had some way back when I made my A290s by hand with a belt sander. I tried the 4001 on them and ended up with the BMS4590, but DSPd not using the passive xover. I may make another pair now I have the CNC just for fun, and sell them on if I don't find a use for them as I'm now sold on MEHs, even though I'm moving back to the house where I used the A290s nearly 20y later. I went MEH because the current apartment is narrow, but the house has a huge room and sidewall reflections are well outside Haas.
I also liked the JBL2446 with Radian Als in them on the A290. Still have the drivers and a set of 2445 I forgot I even owned. And some Emilar EC320s again with Radians.
And they all needed EQ to sound their best.
@Brett : If you have a CNC and are thinking of making more A290 style horns then I’d suggest buying the CAD files from Joseph Crowe and making his ES290 horns which are an evolution of the A290.
https://josephcrowe.com/products/es-290-biradial-wood-horn
https://josephcrowe.com/products/es-290-biradial-wood-horn
I can't see much point in running any horn + CD combo without eq of any sort. While i prefer passive eq via xover, I do use dsp to quickly determine whether a CD and/or WG has potential to make good music. Getting that 1 - 10 k range limear is critical to a correct and coherent sounding setup. Of course there are mismatched combinations of components that just don't sound right with any amount of FR correction but to dismiss a CD + WG combo without corrective EQ is unfair IMO. Its sort of like those Lowther guys that think anything extra in the signal path is evil, yet they listen contently to fullrange drivers that have all sorts of nasty resonances and breakup modes which make me want to run straight out of the room.
I'm aware of them and thought about making some for myself to make surrounds with, but I'm going back to raw 2ch. A friend was interested in them at one stage too but his missus thought they were too large so we have another plan.
Last edited: